SUMMARY Each year, there are over 10 million non-surgical hospitalizations in the United States which trigger approximately 1 in 3 venous thromboembolism (VTE) events ? or about 200,000 VTEs per year. National quality measures and professional societies recommend assessing VTE risk at admission and provide appropriate VTE prevention measures which include ambulation, sequential compression devices, and/or prophylactic dose anticoagulant medications (i.e. enoxaparin). In addition to reducing VTE, pharmacologic VTE prophylaxis also increases bleeding risk. Over the past few years there have been divergent pressures on clinicians regarding VTE prevention; on the one hand physicians are encouraged to assess VTE risk in non- surgical patients and provide VTE prophylaxis, and on the other a growing body of evidence suggests that the benefits of deceases in VTE may be outweighed by increased bleeding. Prior research of hospital-acquired (HA) VTE and bleeding has been hampered by the fact that it is a rare complication of a common event (hospitalization). Tens of thousands of hospitalizations need to be evaluated to obtain sufficient numbers of events to characterize who suffers HA-VTE and HA-bleeding. Prior studies have had to rely on time-consuming chart abstraction of thousands of hospitalizations or rely only on administrative data to study HA-VTE or HA- bleeding. With the introduction of the electronic health record, we can now assess tens of thousands of non- surgical admissions and determine who is at risk for VTE and bleeding. The aims of this proposal are threefold, at two diverse institutions (The University of Vermont and the University of Washington): (i) to develop risk models for HA-VTE and HA-bleeding (ii) to validate risk models for HA-VTE and HA-bleeding, and (iii) to determine the incidence of and risk factors for post-discharge VTE and bleeding. With these aims, we will make healthcare safer by allowing providers to tailor VTE prevention strategies to those at highest risk of VTE and lowest risk of bleeding. Our innovation lies in leveraging the potential of the electronic health record to efficiently study HA-VTE and HA-bleeding events in over 140,000 admissions, and in advancing clinical practice by developing validated risk models to allow providers to efficiently and rationally determine the risk:benefit of pharmacologic VTE prophylaxis at the time of admission and begin to understand the VTE and bleeding risks at discharge.